Artificial Intelligence in Neuroradiology: A Review of Current Topics and Competition Challenges.

There is an expanding body of literature that describes the application of deep learning and other machine learning and artificial intelligence methods with potential relevance to neuroradiology practice. In this article, we performed a literature review to identify recent developments on the topics of artificial intelligence in neuroradiology, with particular emphasis on large datasets and large-scale algorithm assessments, such as those used in imaging AI competition challenges. Numerous applications relevant to ischemic stroke, intracranial hemorrhage, brain tumors, demyelinating disease, and neurodegenerative/neurocognitive disorders were discussed. The potential applications of these methods to spinal fractures, scoliosis grading, head and neck oncology, and vascular imaging were also reviewed. The AI applications examined perform a variety of tasks, including localization, segmentation, longitudinal monitoring, diagnostic classification, and prognostication. While research on this topic is ongoing, several applications have been cleared for clinical use and have the potential to augment the accuracy or efficiency of neuroradiologists.

The added value of a second read by pediatric radiologists for outside skeletal surveys.

BACKGROUND: Fractures are the second most common finding in non-accidental trauma after cutaneous signs. Interpreting skeletal surveys could be challenging as some fractures are subtle and due to anatomical variations that can mimic injuries. OBJECTIVE: To determine the effect of a second read by a pediatric radiologist of skeletal surveys for suspected non-accidental trauma initially read at referring hospitals by general radiologists. MATERIALS AND METHODS: In 2016 and 2017, we identified all patients referred to our children's hospital with previous surveys performed and read at a community hospital by an outside radiologist. We excluded patients older than 3 years and studies performed at a children's hospital. The surveys were reviewed by a pediatric radiologist with the printed outside report available. Surveys with disagreement between outside read and pediatric radiologist read were reviewed by a second pediatric radiologist. A disagreement in the second read included only definite discrepant findings agreed upon by both pediatric radiologists. The Fisher exact test was performed to compare the ratio of discrepancies between readers in normal and abnormal surveys. RESULTS: Two hundred twenty-five surveys were performed (120 male) at 62 referring hospitals, with a mean patient age of 10.5 months (range: 5 days-3 years). The outside read identified fractures in 104/225 (46.2%) surveys. Thirty-seven of the 225 (16.4%) contained discrepancies in interpretation (n=111). Most of these disagreements (29/37, 78.4%) resulted in a significant change in the report. There was a significant (P<0.0001) difference between disagreement rate in outside read negative (4/111, 3.2%) and positive surveys (34/104, 31.7%). The second read identified additional fractures in 22/225 (9.8%) of the surveys and disagreed with first-read fractures in 17/256 (7.6%). Four of 19 (21.1%) classic metaphyseal lesions diagnosed by the outside read were normal variants; 18 classic metaphyseal lesions were missed by the outside read. CONCLUSIONS: This study supports second reads by pediatric radiologists of skeletal surveys for non-accidental trauma.